History of AI Flashcards
Describe 3 advances in the early era of AI.
1952: Samuel’s checkers program learned to play at strong amateur level.
1955: Newell & Simon’s Logic Theorist proved theorems in Principia Mathematica.
1956: Foundation principle of AI defined by McCarthy, Minsky, and Shannon: “Every aspect of learning or any other feature of intelligence can be so precisely descried that a machine can be made to simulate it”.
Name 2 problems with the early era of AI.
Limited computation as the search space grew exponentially, outpacing hardware.
Limited information because of the complexity of AI problems.
Name 2 contributions from the early era of AI.
List, garbage collection, time-sharing (John McCarthy)
Key paradigm: separate modelling (declarative) and algorithms (procedural).
Many knowledge-based systems were developed in the 70s-80s. What is an expert system?
They gain specific domain knowledge from experts in the form of rules, e.g.
if [premises] then [conclusion]
Name and describe a knowledge-based system developed in the 70s/80s.
DENDRAL: infer molecular structure from mass spectrometry.
MYCIN: diagnose blood infections, recommend antibodies.
XCON: concert customer orders into parts specification.
Name 2 contributions in AI from the 70s/80s.
First real application that impacted industry created.
Knowledge helped curb the exponential growth.
Name 2 problems with the developments made in AI in the 70s/80s.
Knowledge is not deterministic rules, uncertainty needs to be modelled.
Requires considerable manual effort to create rules, and thus is hard to maintain.
Name 2 developments of more modern AI (80s/90s).
Probability - Pearl (1988) promotes Bayesian networks in AI to model uncertainty (based on Bayes rule from 1700s).
Machine Learning - Vapnik (1995) invented support vector machines to tune parameters (based on statistical models in early 1900s).
